Evaluation of a Machine Learning-Based Prognostic Model for Unrelated Hematopoietic Cell Transplantation Donor Selection.

The survival of patients undergoing hematopoietic cell transplantation (HCT) from unrelated donors for acute leukemia exhibits considerable variation, even after stringent genetic matching. To improve the donor selection process, we attempted to create an algorithm to quantify the likelihood of survival to 5 years after unrelated donor HCT for acute leukemia, based on the clinical characteristics of the donor selected. All standard clinical variables were included in the model, which also included average leukocyte telomere length of the donor based on its association with recipient survival in severe aplastic anemia, and links to multiple malignancies. We developed a multivariate classifier that assigned a Preferred or NotPreferred label to each prospective donor based on the survival of the recipient. In a previous analysis using a resampling method, recipients with donors labeled Preferred experienced clinically compelling better survival compared with those labeled NotPreferred by the test. However, in a pivotal validation study in an independent cohort of 522 patients, the overall survival of the Preferred and NotPreferred donor groups was not significantly different. Although machine learning approaches have successfully modeled other biological phenomena and have led to accurate predictive models, our attempt to predict HCT outcomes after unrelated donor transplantation was not successful.

Development of an Unrelated Donor Selection Score Predictive of Survival after HCT: Donor Age Matters Most.

Donor factors, in addition to HLA matching status, have been associated with recipient survival in unrelated donor (URD) hematopoietic cell transplantation (HCT); however, there is no hierarchical algorithm that weights the characteristics of individual donors against each other in a quantitative manner to facilitate donor selection. The goal of this study was to develop and validate a donor selection score that prioritizes donor characteristics associated with better survival in 8/8 HLA-matched URDs. Two separate patient/donor cohorts, the first receiving HCT between 1999 and 2011 (n = 5952, c1), and the second between 2012 and 2014 (n = 4510, c2) were included in the analysis. Both cohorts were randomly spilt, 2:1, into training and testing sets. Despite studying over 10,000 URD transplants, we were unable to validate a donor selection score. The only donor characteristic associated with better survival was younger age, with 2-year survival being 3% better when a donor 10 years younger is selected. These results support previous studies suggesting prioritization of a younger 8/8 HLA-matched donor. This large dataset also shows that none of the other donor clinical factors tested were reproducibly associated with survival, and hence flexibility in selecting URDs based on other characteristics is justified. These data support a simplified URD selection process and have significant implications for URD registries.

MHC Class I Chain-Related Gene A (MICA) Donor-Recipient Mismatches and MICA-129 Polymorphism in Unrelated Donor Hematopoietic Cell Transplantations Has No Impact on Outcomes in Acute Lymphoblastic Leukemia, Acute Myeloid Leukemia, or Myelodysplastic Syndrome: A Center for International Blood and Marrow Transplant Research Study.

Single-center studies have previously reported associations of MHC Class I Chain-Related Gene A (MICA) polymorphisms and donor-recipient MICA mismatching with graft-versus-host disease (GVHD) after unrelated donor hematopoietic cell transplantation (HCT). In this study, we investigated the association of MICA polymorphism (MICA-129, MM versus MV versus VV) and MICA mismatches after HCT with 10/10 HLA-matched (n = 552) or 9/10 (n = 161) unrelated donors. Included were adult patients with a first unrelated bone marrow or peripheral blood HCT for acute lymphoblastic leukemia, acute myeloid leukemia, or myelodysplastic syndrome that were reported to the Center for International Blood and Marrow Transplant Research between 1999 and 2011. Our results showed that neither MICA mismatch nor MICA-129 polymorphism were associated with any transplantation outcome (P < .01), with the exception of a higher relapse in recipients of MICA-mismatched HLA 10/10 donors (hazard ratio [HR], 1.7; P = .003). There was a suggestion of association between MICA mismatches and a higher risk of acute GVHD grades II to IV (HR, 1.4; P = .013) There were no significant interactions between MICA mismatches and HLA matching (9/10 versus 10/10). In conclusion, the findings in this cohort did not confirm prior studies reporting that MICA polymorphism and MICA mismatches were associated with HCT outcomes.

Analysis of a Genetic Polymorphism in the Costimulatory Molecule TNFSF4 with Hematopoietic Stem Cell Transplant Outcomes.

Despite stringent procedures to secure the best HLA matching between donors and recipients, life-threatening complications continue to occur after hematopoietic stem cell transplantation (HSCT). Studying single nucleotide polymorphism (SNP) in genes encoding costimulatory molecules could help identify patients at risk for post-HSCT complications. In a stepwise approach we selected SNPs in key costimulatory molecules including CD274, CD40, CD154, CD28, and TNFSF4 and systematically analyzed their association with post-HSCT outcomes. Our discovery cohort analysis of 1157 HLA-A, -B, -C, -DRB1, and -DQB1 matched cases found that patients with donors homozygous for the C variant of rs10912564 in TNFSF4 (48%) had better disease-free survival (P = .029) and overall survival (P = .009) with less treatment-related mortality (P = .006). Our data demonstrate the TNFSF4C variant had a higher affinity for the nuclear transcription factor Myb and increased percentage of TNFSF4-positive B cells after stimulation compared with CT or TT genotypes. However, these associations were not validated in a more recent cohort, potentially because of changes in standard of practice or absence of a true association. Given the discovery cohort, functional data, and importance of TNFSF4 in infection clearance, TNFSF4C may associate with outcomes and warrants future studies.

Toll-like receptor polymorphisms in allogeneic hematopoietic cell transplantation.

To assess the impact of the genetic variation in toll-like receptors (TLRs) on outcome after allogeneic myeloablative conditioning hematopoietic cell transplantation (HCT), we investigated 29 single nucleotide polymorphisms across 10 TLRs in 816 patients and donors. Only donor genotype of TLR8 rs3764879, which is located on the X chromosome, was significantly associated with outcome at the Bonferroni-corrected level P ≤ .001. Male hemizygosity and female homozygosity for the minor allele were significantly associated with disease-free survival (hazard ratio [HR], 1.47 [95% confidence interval {CI}, 1.16 to 1.85]; P = .001). Further analysis stratified by donor sex due to confounding by sex was suggestive for associations with overall survival (male donor: HR, 1.41 [95% CI, 1.09 to 1.83], P = .010; female donor: HR, 2.78 [95% CI, 1.43 to 5.41], P = .003), disease-free survival (male donor: HR, 1.45 [95% CI, 1.12 to 1.87], P = .005; female donor: HR, 2.34 [95% CI, 1.18 to 4.65], P = .015), and treatment-related mortality (male donor: HR, 1.49 [95% CI, 1.09 to 2.04], P = .012; female donor: HR, 3.12 [95% CI, 1.44 to 6.74], P = .004). In conclusion, our findings suggest that the minor allele of TLR8 rs3764879 of the donor is associated with outcome after myeloablative conditioned allogeneic HCT.